Bitcoin’s consensus rules had not changed since Taproot activated in November 2021. That is the longest quiet stretch in the network’s history.
BIP-110 tried to end it. The proposal set its own activation threshold at 55% of miners.
It got 2.53%.
What Happened at Block 961,632
The mandatory signaling window opened at block 961,632 on August 8, 2026.
Nodes running BIP-110 software began rejecting blocks that did not signal support. Almost no blocks did, so those nodes rejected the chain that nearly all miners were building.
That produced a live minority fork. It made two blocks in roughly eight hours, then stalled, while the main network kept its normal pace and pulled dozens of blocks ahead.
Roughly 99.85% of Bitcoin’s hashpower stayed on the original chain.
The proposal’s formal name is the Reduced Data Temporary Softfork. Despite social media descriptions to the contrary, it is not a quantum security upgrade. Its purpose was to temporarily restrict arbitrary data embedded in transactions.
The Signaling Numbers Do Not Match, and That Is Fine
You will see three different figures quoted for miner support. All three appear in credible coverage.
One analysis counted roughly 38 of more than 9,000 blocks signaling since May 1, which is about 0.42%. Another counted 51 of 2,016 blocks in the immediately preceding difficulty period, which is 2.53%. A third reported 2.59% shortly before the window opened.
These are not contradictory. They measure different windows. The long-run figure is low because support was near zero for months. The recent figure is higher because a handful of blocks signaled as the deadline approached.
Whichever you use, the gap to 55% is enormous. There was never a version of this that activated.
| Measure | Figure | Window |
| Required for activation | 55% | Set by the proposal itself |
| Long-run signaling | about 0.42% | Since May 1, 2026 |
| Recent signaling | 2.53% | Prior 2,016-block period |
| Pre-window reading | about 2.59% | Early August 2026 |
| Hashpower on main chain | about 99.85% | At the split |
The Replay Problem Nobody Designed For
BIP-110 shipped without replay protection. Its authors did not include any because they were not building a new network. They were tightening rules on the existing one.
That logic held right up until consensus failed and a split happened anyway.
Replay protection stops a transaction broadcast on one chain from being valid on the other. Bitcoin Cash implemented it in 2017 precisely to avoid this problem. BIP-110 has none.
So during the minority chain’s operating window, a transaction signed on one chain could in principle be rebroadcast on the other, moving coins the sender never intended to move.
The Specific Trap
Developer Kevin Loaec flagged the practical risk on August 6, before the window opened.
The setup works like this. A holder with 10 BTC before the split controls 10 coins on each chain afterward. The second balance looks like free money, and a buyer may offer to purchase the fork coins.
Because both networks can initially recognize the same signed transaction, that buyer can copy the transaction used to send the fork coins and broadcast it on the main Bitcoin network.
The seller then loses real BTC as well as the fork coins. The buyer gets both.
The safe response is the boring one. Leaving coins untouched carries no replay exposure at all.
Who Was Actually Exposed
This distinction got lost in a lot of coverage, so it is worth stating plainly.
Coins held on major exchanges were not affected. Exchanges operate on the main chain, which carries the overwhelming majority of hashpower and economic value.
Spot bitcoin ETF holders were not affected either. The funds track the dominant chain, and the operational work sits with the custodian rather than the investor.
The exposure was concentrated among self-custody holders running BIP-110 enforcing software, and among anyone who tried to sell fork coins during the window.
For everyone else, sending and receiving bitcoin worked normally throughout.
The Opposition Was the Story
Michael Saylor published a 110-point essay opposing the proposal, calling it extremely dangerous.
His argument was about precedent rather than spam. Rejecting valid, fee-paying transactions establishes that some transactions can be refused on the basis of their content, and that principle can be extended to any class of Bitcoin activity later.
Blockstream’s Adam Back took a similar line, arguing Bitcoin should not police other systems through code, and that forking off remains the recourse for anyone who disagrees.
Saylor has been a lightning rod in these debates before, including when he downplayed self-custody. On this one, the miners ended up where he did.
What the Failure Proved
BIP-110 will be remembered less for what it proposed than for what its collapse demonstrated.
Changing Bitcoin now requires alignment across miners, node operators, developers, exchanges and large holders simultaneously. A proposal can be technically coherent, have a working implementation, and reach a live signaling window, and still get 2.53%.
That is either reassuring or paralyzing depending on your view of what Bitcoin should become. The block size war produced the same lesson a decade ago, a period Optimisus revisited when Vitalik Buterin reflected on it.
One date needs flagging. Some coverage described the proposal as officially dead on August 1, a week before the signaling window opened. That framing does not square with the block-height timeline, and readers should treat August 8 as the operative date.
What Comes Next
The signaling window runs to block 963,647. It is arithmetically over regardless.
A separate and much larger event follows. A hard fork called eCash is targeted at block 964,000, around August 21, and unlike BIP-110 it is designed to split.
The covenant debate is the longer-running one. OP_CTV and OP_CAT would make Bitcoin substantially more programmable. OP_CAT reached complete specification status on March 1, 2026 with no mainnet activation parameters, and OP_CTV has a published activation client aimed at 2027.
Longer-horizon threats sit behind all of it, including the quantum question raised when Tether’s chief executive warned that inactive bitcoin wallets could eventually be exposed.
Neither has broad consensus. After watching a proposal draw 2.53%, the people behind both are recalculating what consensus would even require.
Bitcoin traded near $63,000 through the split. The market treated the whole episode as a non-event, which was the correct read.
Sources
- Crypto.news, BIP-110 is dead, what Bitcoin’s failed anti-spam fork reveals about governance in 2026 — https://crypto.news/bip-110-bitcoin-fork-failed-governance-2026/
- Crypto.news, Bitcoin BIP-110 fork could expose holders to replay attacks — https://crypto.news/bitcoin-bip-110-fork-expose-holders-to-replay-attacks/
- Blockonomi, Bitcoin BIP-110 fork splits chain as main BTC network pulls ahead, August 9, 2026 — https://blockonomi.com/bitcoin-bip-110-fork-splits-chain-as-main-btc-network-pulls-ahead/
- Yellow, Bitcoin nears a BIP-110 fork with just 2.59% miner support — https://yellow.com/news/bitcoin-bip-110-fork-miner-support
- AMINA Bank research, Bitcoin fork August 2026, BIP-110, eCash, covenants and the quantum clock — https://aminagroup.com/research/bitcoin-fork-august-2026-bip-110-ecash-covenants-and-the-quantum-clock/
- Tech Times, Bitcoin’s BIP-110 soft fork is dead, what a minority chain split means for holdings — https://www.techtimes.com/articles/322684/20260802/bitcoins-bip-110-soft-fork-dead-what-minority-chain-split-means-your-holdings.htm
This is not financial advice.
Optimisus covers crypto and technology news for readers who want the detail behind the headline.
